Google’s AI news headlines stay despite daily mistakes

A large, glossy Google logo at center surrounded by a collage of photo thumbnails and blank newspaper rectangles, some misaligned and glitching, on a bright white background with splashes of Google red, blue, yellow, and green, medium close-up framing with crisp editorial lighting, no text visible

Google told reporters this week that its AI-generated headlines in the Discover news feed are now a permanent feature. The company says the system performs well for user satisfaction. But journalists and readers continue to find false claims and misleading summaries almost daily.

False Claims and Confused Stories

According to The Verge, Google’s AI recently wrote that the US reversed a foreign drone ban. The headline linked to a PCMag story that explained the exact opposite. PCMag specifically warned readers that such headlines were misleading.

The AI also confused different companies and products. One headline claimed new 3D technology came from a company called Leia. But the linked story was about a different firm called Visual Semiconductor. Another headline said the ASUS ROG Ally X arrived in January. The device actually launched in 2024.

Jim Fisher, who wrote the PCMag drone story, said the AI headlines make him feel icky. He told The Verge that Google should use the headlines that humans wrote. If Google needs summaries, it can use the ones that publications already submit.

How the System Works

Google calls these AI creations trending topics. Each one presents itself as a news story from a specific publication. The system uses the publication’s images and links to its articles. But it replaces the original headline with an AI-generated version.

Google spokesperson Jennifer Kutz said the feature helps people explore topics covered by multiple sites. She said the AI headline reflects information across a range of sites. It is not a rewrite of one article headline.

Impact on Publishers

The system frustrates journalists who craft headlines to attract readers. One Verge story about RGB stripe OLED monitors became New OLED Gaming Monitors Debut. A piece about immersive 3D video turned into Lego Smart Play launches March 1.

Google also serves human-written clickbait without changes. One headline promised a free Star Wars game download. The actual giveaway was a single code for UK residents only.

The feature now appears beyond the Discover feed. Some users see AI headlines as push notifications. Tapping them opens a Google Gemini chatbot that tries to summarize recent news.

Total
0
Shares
Previous Post
Close-up portrait crop of an anonymous woman’s face from nose to cheekbones, left side dry and expressionless and right side with a single glossy tear added, faint White House facade softly blurred in the background, bright high-key look with cool blues against warm amber highlights, crisp editorial montage style, centered and symmetrical, no text

White House adds fake tears to arrest photo with AI

Next Post
Medium close-up of President Donald Trump in a winter parka striding across bright blue Arctic ice toward towering Greenland-style ice cliffs, a small penguin at his side holding a tiny American flag, oddly identical human and penguin footprints trailing behind, icy cyan and deep navy contrasted with warm orange highlights, dramatic sunrise sky, high clarity, central composition, no text or logos

White House shares AI image with penguin that doesn’t belong in Greenland

Related Posts